Separate waiting for the response, discovering the image, downloading it and rendering it.
Name the element before investigating it
The largest visible content element can change between layouts. Confirm which element the performance report identifies on the page and viewport you are testing. Save the URL, viewport and a screenshot. If the largest element is text rather than the image you expected, an image-only investigation would miss the actual issue.
Record the four parts of the wait
Google's LCP guidance separates time to first byte, resource load delay, load duration and render delay. Use that breakdown as four columns in your worksheet. A short download paired with a late render calls for a different investigation from a large file that begins downloading promptly.
Capture a comparable request trace
Run a fresh load under documented conditions. Save the document timing, the image request start and completion, and the reported LCP time. Note redirects or resource discovery that occurs only after other work. Collect several comparable runs so a single unusually fast or slow request does not become the entire explanation.
Propose one testable change
If the image is oversized, investigate responsive sizing. If discovery is delayed, ask the developer to examine how the element is exposed in markup. If rendering waits after download, ask what keeps it hidden or blocked. State the expected effect before making the change so you can tell whether the result supports your hypothesis.
Send a report with a stopping condition
Include the original trace, proposed change, new trace and functional checks. Do not promise a ranking increase or a fixed speed gain from the change. If the targeted delay remains, restore the configuration or investigate the next supported explanation. Keep the opening content usable while refining delivery.
Put it into practice
- Confirm the actual LCP element and viewport.
- Record the four timing components.
- Keep several comparable traces.
- Change one suspected cause.
- Compare timing and functional behavior afterward.
AI-assisted educational content prepared for the Velonic resource library. How these guides are prepared.